Research And Implementation Of Face Recognition Algorithm Based On Deep Learning And Graph Theory

With the development of human society,the identity information are increasingly valued by people.The faces carries much information,people can achieve the identity of the people by the human face.After several decades of development,the face recognition technology has been applied in many aspects,such as entrance guard,monitoring,criminal investigation and so on.These cases mean that it has significant meaning to research the face recognition algorithm and improve the accuracy and efficiency of face recognition.The paper's research starts from the following contents:(1)The development and current situation of the face recognition technology both at home and abroad are introduced.Through the research of face recognition technology at home and abroad,the key technology of face recognition is understood.And then through the research of deep learning,the deep learning can combine with the graph theory to solve the problem of face recognition.And achieved good results in this way.(2)The face recognition algorithm based on graph and ASM model is designed and achieved.The ASM model are trained firstly in this algorithm,then the shape model is established.In the process of training,the samples are collected firstly,then the key points of the face images are recorded manually,the shape vectors of the training set are constructed and the shape is normalized.Then the local features of the key points are constructed.In the process of identifying algorithm,the key points of the faces in images are detected by the ASM model.then the face graphs are constructed by neighborhood algorithm and the Gabor Jets around the key points are computed.Then the similarities between the face graph which will be recognized and the model face graphs are computed to achieve the graph matching.Finally the corresponding face can be recognized.The experiments on the ORL face database and the AR face database show that the algorithm achieves the ideal effect.(3)The face recognition algorithm based on deep learning and graph is designed and achieved.In this recognition algorithm,the original images are preprocessed firstly,then the size and the gray level are normalized,these pre-processed images are used to train the convolution neural network,the image to be recognized is preprocessed in the same way and then input it into the trained neural network.Then the key points of the face in images can be detected.The face graphs are constructed and then the graph matching operation will be completed to achieve the identification.Experiments on the ORL face database and the AR face database show that the algorithm also achieves the ideal effect.